bgpd: evpn pip handle svi ip route

By default announct Self Type-2 routes with
system IP as nexthop and system MAC as
nexthop.

An API to check type-2 is self route via
checking ipv4/ipv6 address from connected interfaces list.

An API to extract RMAC and nexthop for type-2
routes based on advertise-svi-ip knob is enabled.

When advertise-pip is enabled/disabled, trigger type-2
route update. For self type-2 routes to use
anycast or individual (rmac, nexthop) addresses.

Ticket:CM-26190
Reviewed By:
Testing Done:

Enable 'advertise-svi-ip' knob in bgp default instance.
the vrf instance svi ip is advertised with nexthop
as default instance router-id and RMAC as system MAC.

Signed-off-by: Chirag Shah <chirag@cumulusnetworks.com>

7 files changed, 105 insertions, 4 deletions

diff --git a/bgpd/bgp_evpn.c b/bgpd/bgp_evpn.c
index 49b2b7d1ed..c31073cc8f 100644
--- a/bgpd/bgp_evpn.c
+++ b/bgpd/bgp_evpn.c
@@ -494,6 +494,39 @@ static void unmap_vni_from_rt(struct bgp *bgp, struct bgpevpn *vpn,
 	}
 }
 
+static void bgp_evpn_get_rmac_nexthop(struct bgpevpn *vpn,
+				      struct prefix_evpn *p,
+				      struct attr *attr, uint8_t flags)
+{
+	struct bgp *bgp_vrf = vpn->bgp_vrf;
+
+	memset(&attr->rmac, 0, sizeof(struct ethaddr));
+	if (!bgp_vrf)
+		return;
+
+	if (p->prefix.route_type == BGP_EVPN_MAC_IP_ROUTE) {
+		/* Copy sys (pip) RMAC and PIP IP as nexthop
+		 * in case of route is self MAC-IP,
+		 * advertise-pip and advertise-svi-ip features
+		 * are enabled.
+		 * Otherwise, for all host MAC-IP route's
+		 * copy anycast RMAC
+		 */
+		if (CHECK_FLAG(flags, BGP_EVPN_MACIP_TYPE_SVI_IP)
+		    && bgp_evpn_is_svi_macip_enabled(vpn)) {
+			/* copy sys rmac */
+			memcpy(&attr->rmac, &bgp_vrf->evpn_info->pip_rmac,
+			       ETH_ALEN);
+			if (bgp_vrf->evpn_info->advertise_pip &&
+			    bgp_vrf->evpn_info->is_anycast_mac) {
+				attr->nexthop = bgp_vrf->evpn_info->pip_ip;
+				attr->mp_nexthop_global_in =
+					bgp_vrf->evpn_info->pip_ip;
+			}
+		} else
+			memcpy(&attr->rmac, &bgp_vrf->rmac, ETH_ALEN);
+	}
+}
 /*
  * Create RT extended community automatically from passed information:
  * of the form AS:VNI.
@@ -1685,6 +1718,9 @@ static int update_evpn_route_entry(struct bgp *bgp, struct bgpevpn *vpn,
 
 		memcpy(&tmp_pi->extra->label, label, sizeof(label));
 		tmp_pi->extra->num_labels = num_labels;
+		/* Mark route as self type-2 route */
+		if (flags && CHECK_FLAG(flags, ZEBRA_MACIP_TYPE_SVI_IP))
+			tmp_pi->extra->af_flags = BGP_EVPN_MACIP_TYPE_SVI_IP;
 		bgp_path_info_add(rn, tmp_pi);
 	} else {
 		tmp_pi = local_pi;
@@ -1828,8 +1864,29 @@ static int update_evpn_route(struct bgp *bgp, struct bgpevpn *vpn,
 	}
 
 	/* router mac is only needed for type-2 routes here. */
-	if (p->prefix.route_type == BGP_EVPN_MAC_IP_ROUTE)
-		bgpevpn_get_rmac(vpn, &attr.rmac);
+	if (p->prefix.route_type == BGP_EVPN_MAC_IP_ROUTE) {
+		uint8_t af_flags = 0;
+
+		if (CHECK_FLAG(flags, ZEBRA_MACIP_TYPE_SVI_IP))
+			SET_FLAG(af_flags, BGP_EVPN_MACIP_TYPE_SVI_IP);
+
+		bgp_evpn_get_rmac_nexthop(vpn, p, &attr, af_flags);
+
+		if (bgp_debug_zebra(NULL)) {
+			char buf[ETHER_ADDR_STRLEN];
+			char buf1[PREFIX_STRLEN];
+
+			zlog_debug("VRF %s vni %u type-2 route evp %s RMAC %s nexthop %s",
+				   vpn->bgp_vrf ?
+				   vrf_id_to_name(vpn->bgp_vrf->vrf_id) : " ",
+				   vpn->vni,
+				   prefix2str(p, buf1, sizeof(buf1)),
+				   prefix_mac2str(&attr.rmac, buf,
+						  sizeof(buf)),
+				   inet_ntoa(attr.mp_nexthop_global_in));
+		}
+	}
+
 	vni2label(vpn->vni, &(attr.label));
 
 	/* Include L3 VNI related RTs and RMAC for type-2 routes, if they're
@@ -2104,7 +2161,8 @@ static int update_all_type2_routes(struct bgp *bgp, struct bgpevpn *vpn)
 		attr.nexthop = vpn->originator_ip;
 		attr.mp_nexthop_global_in = vpn->originator_ip;
 		attr.mp_nexthop_len = BGP_ATTR_NHLEN_IPV4;
-		bgpevpn_get_rmac(vpn, &attr.rmac);
+		bgp_evpn_get_rmac_nexthop(vpn, evp, &attr,
+					  tmp_pi->extra->af_flags);
 
 		if (evpn_route_is_sticky(bgp, rn))
 			attr.sticky = 1;
@@ -2114,6 +2172,19 @@ static int update_all_type2_routes(struct bgp *bgp, struct bgpevpn *vpn)
 				attr.router_flag = 1;
 		}
 
+		if (bgp_debug_zebra(NULL)) {
+			char buf[ETHER_ADDR_STRLEN];
+			char buf1[PREFIX_STRLEN];
+
+			zlog_debug("VRF %s vni %u evp %s RMAC %s nexthop %s",
+				   vpn->bgp_vrf ?
+				   vrf_id_to_name(vpn->bgp_vrf->vrf_id) : " ",
+				   vpn->vni,
+				   prefix2str(evp, buf1, sizeof(buf1)),
+				   prefix_mac2str(&attr.rmac, buf, sizeof(buf)),
+				   inet_ntoa(attr.mp_nexthop_global_in));
+		}
+
 		/* Add L3 VNI RTs and RMAC for non IPv6 link-local if
 		 * using L3 VNI for type-2 routes also.
 		 */
@@ -2301,7 +2372,7 @@ static int bgp_evpn_vni_flood_mode_get(struct bgp *bgp,
  * situations need the route in the per-VNI table as well as the global
  * table to be updated (as attributes change).
  */
-static int update_routes_for_vni(struct bgp *bgp, struct bgpevpn *vpn)
+int update_routes_for_vni(struct bgp *bgp, struct bgpevpn *vpn)
 {
 	int ret;
 	struct prefix_evpn p;
diff --git a/bgpd/bgp_evpn_private.h b/bgpd/bgp_evpn_private.h
index a0c05e34f8..76cf8b2cd6 100644
--- a/bgpd/bgp_evpn_private.h
+++ b/bgpd/bgp_evpn_private.h
@@ -511,6 +511,16 @@ static inline int is_es_local(struct evpnes *es)
 	return CHECK_FLAG(es->flags, EVPNES_LOCAL) ? 1 : 0;
 }
 
+static inline bool bgp_evpn_is_svi_macip_enabled(struct bgpevpn *vpn)
+{
+	struct bgp *bgp_evpn = NULL;
+
+	bgp_evpn = bgp_get_evpn();
+
+	return (bgp_evpn->evpn_info->advertise_svi_macip ||
+		vpn->advertise_svi_macip);
+}
+
 extern void bgp_evpn_install_uninstall_default_route(struct bgp *bgp_vrf,
 						     afi_t afi, safi_t safi,
 						     bool add);
@@ -553,4 +563,5 @@ extern struct evpnes *bgp_evpn_es_new(struct bgp *bgp, esi_t *esi,
 				      struct ipaddr *originator_ip);
 extern void bgp_evpn_es_free(struct bgp *bgp, struct evpnes *es);
 extern bool bgp_evpn_lookup_l3vni_l2vni_table(vni_t vni);
+extern int update_routes_for_vni(struct bgp *bgp, struct bgpevpn *vpn);
 #endif /* _BGP_EVPN_PRIVATE_H */
diff --git a/bgpd/bgp_evpn_vty.c b/bgpd/bgp_evpn_vty.c
index 882716cf37..5508ab99b4 100644
--- a/bgpd/bgp_evpn_vty.c
+++ b/bgpd/bgp_evpn_vty.c
@@ -3767,7 +3767,17 @@ DEFPY (bgp_evpn_advertise_pip_ip_mac,
 	}
 
 	if (is_evpn_enabled()) {
+		struct listnode *node = NULL;
+		struct bgpevpn *vpn = NULL;
+
 		update_advertise_vrf_routes(bgp_vrf);
+
+		/* Update (svi) type-2 routes */
+		for (ALL_LIST_ELEMENTS_RO(bgp_vrf->l2vnis, node, vpn)) {
+			if (!bgp_evpn_is_svi_macip_enabled(vpn))
+				continue;
+			update_routes_for_vni(bgp_evpn, vpn);
+		}
 	}
 
 	return CMD_SUCCESS;
diff --git a/bgpd/bgp_route.h b/bgpd/bgp_route.h
index a710873ea7..b9f3f3f762 100644
--- a/bgpd/bgp_route.h
+++ b/bgpd/bgp_route.h
@@ -114,6 +114,10 @@ struct bgp_path_info_extra {
 	mpls_label_t label[BGP_MAX_LABELS];
 	uint32_t num_labels;
 
+	/* af specific flags */
+	uint16_t af_flags;
+#define BGP_EVPN_MACIP_TYPE_SVI_IP (1 << 0)
+
 #if ENABLE_BGP_VNC
 	union {
 
diff --git a/lib/zclient.h b/lib/zclient.h
index ebf3e6450f..7adb294a31 100644
--- a/lib/zclient.h
+++ b/lib/zclient.h
@@ -490,6 +490,7 @@ enum zapi_iptable_notify_owner {
 #define ZEBRA_MACIP_TYPE_GW                    0x02 /* gateway (SVI) mac*/
 #define ZEBRA_MACIP_TYPE_ROUTER_FLAG           0x04 /* Router Flag - proxy NA */
 #define ZEBRA_MACIP_TYPE_OVERRIDE_FLAG         0x08 /* Override Flag */
+#define ZEBRA_MACIP_TYPE_SVI_IP                0x10 /* SVI MAC-IP */
 
 enum zebra_neigh_state { ZEBRA_NEIGH_INACTIVE = 0, ZEBRA_NEIGH_ACTIVE = 1 };
 
diff --git a/zebra/zebra_vxlan.c b/zebra/zebra_vxlan.c
index ba441d7a74..b2cb726fac 100644
--- a/zebra/zebra_vxlan.c
+++ b/zebra/zebra_vxlan.c
@@ -2493,6 +2493,8 @@ static int zvni_neigh_send_add_to_client(vni_t vni, struct ipaddr *ip,
 	/* Set router flag (R-bit) based on local neigh entry add */
 	if (CHECK_FLAG(neigh_flags, ZEBRA_NEIGH_ROUTER_FLAG))
 		SET_FLAG(flags, ZEBRA_MACIP_TYPE_ROUTER_FLAG);
+	if (CHECK_FLAG(neigh_flags, ZEBRA_NEIGH_SVI_IP))
+		SET_FLAG(flags, ZEBRA_MACIP_TYPE_SVI_IP);
 
 	return zvni_macip_send_msg_to_client(vni, macaddr, ip, flags,
 			     seq, ZEBRA_NEIGH_ACTIVE, ZEBRA_MACIP_ADD);
@@ -2813,6 +2815,7 @@ static int zvni_gw_macip_add(struct interface *ifp, zebra_vni_t *zvni,
 					      n->flags, n->loc_seq);
 	} else if (advertise_svi_macip_enabled(zvni)) {
 
+		SET_FLAG(n->flags, ZEBRA_NEIGH_SVI_IP);
 		if (IS_ZEBRA_DEBUG_VXLAN)
 			zlog_debug(
 			"SVI %s(%u) L2-VNI %u, sending SVI MAC %s IP %s add to BGP with flags 0x%x",
diff --git a/zebra/zebra_vxlan_private.h b/zebra/zebra_vxlan_private.h
index 784dcdb9d8..f0bda44fb9 100644
--- a/zebra/zebra_vxlan_private.h
+++ b/zebra/zebra_vxlan_private.h
@@ -366,6 +366,7 @@ struct zebra_neigh_t_ {
 #define ZEBRA_NEIGH_DEF_GW    0x08
 #define ZEBRA_NEIGH_ROUTER_FLAG 0x10
 #define ZEBRA_NEIGH_DUPLICATE 0x20
+#define ZEBRA_NEIGH_SVI_IP 0x40
 
 	enum zebra_neigh_state state;